Infrared Communication for Light Shows

[carlk]

Since NRZI is geared more for a continuous data stream, your last bit will determine weather the IR LED is on or off at the end of the data packet.
Leaving it on until the start of the next packet may not be a problem though.
As for a special hat, at 2400 baud with the 25ms rest you should be able to send 12 channels of data out and just have one hat use the second 6 channels.
Carl

[ErnieHorning]

You can do like LOR and everyone have its own address that it looks for but address 255 is global so everyone responds.

[LabRat]

I can't see it ever approaching a full byte for addresses (that's 1/2 of a DMX univers), so perhaps a protocol that uses part of the command byte for addressing, and the rest to indicate the command.

(Ref: see PC to TRIKS-C protocol for an example)

You can do like LOR and everyone have its own address that it looks for but address 255 is global so everyone responds.

[lightman]

New Update to IR Communication....

I successfully achieved 4800 Baud serial data stream to my headset today by replacing the IR Detector with a new Vishay TSOP34156 unit ($1.12/each from Mouser). I am running my VIXEN show at 50 msec update rate with 2 GECE bulbs that are keeping up with room to spare! Fades are cleaner and of course the sparkles have never looked so good.

As you may know... I had acheived 2400 Baud with a slower IR detector running at 38 kHz carrier, but was limited to 100 msec light effect updates due to exceeding the processing window timeline. The new 34156 detector supports a faster carrier frequency at 56 kHz and allows a minimum of 10 carrier cycles per "1" or "0" to maintain a 50% duty cycle.

This means I can easily transmit the 8 channels (10 bits/channel ...with start and stop bits) of lighting control via RS 232 serial stream at 4800 bits/sec within a 16 msec window (Previously it took 33 msec). The stop bits that were flaky before are now rock solid allowing the serial receiver in the SX processor to cleanly sample data stream. The IR transmitter is no longer extending into the next 50 msec period and is able to keep up with vixen data burst rate with room to spare. True Vixen rate of 50 msec periods has been obtained.

In addition to the new IR Detector, the IR transmitter and headset receiver code was quickly updated for the faster carrier with just a few tweaks of the code. The IR LED Emitters have no trouble keeping up the faster modulation.

What are the downfalls?.....Due to the faster detection, the IR Detector has given up some of its ability to filter out stray flourescent light radiation. But this is not a problem at night with typical light shows. I know this because my small "over the bench flourescents" did cause some spurious flickers of GECE bulbs.... but when I turned one of the 4 flourescents off, the issue went away. This only occurred when I put the headset right under the lights (about 2 ft away). This really is not an issue with my final venue, but thought you would like to know.

Full distance/range testing still needs to be done, but doesn't look bad. I am still running the lower wattage IR LED emitters (qty 2 of the 15 mW units) and have not tried the new 160 mW units that just arrived. In lab i ran a quick and dirty range test that seem to have about the same performance as before .... I took unit into hallway and it still responded. Need to run previous range tests with same set up... that will be later.

In summary, 4800 Baud communication works with full 8 channels. This opens up possibilty of sending even more channels to the remote headset especially if I start compacting the GECE 4 bit hues into 8 bit byte packets.... I can double the effective GECE channel count. Minor degradation of floursecent filtering should prove no issue with our use in light shows. Preliminary distance tests prove more than adequate. Need to boost IR transmitter to full power now and see what system will do.

I am jazzed!!!!

Regards,
Lightman

[David_AVD]

Minor degradation of floursecent filtering should prove no issue with our use in light shows.

I wonder if the PWM rate of some LED lighting will have any adverse effects on the IR detector?

[lightman]

I wonder if the PWM rate of some LED lighting will have any adverse effects on the IR detector?

In theory no.. for couple of reasons...

1) They aren't flickering or PWM'ed at 56 kHz ( the modulated carrier freq of the detector)
2) Their wavelength or freq band is not in the detecting range of the IR detector (with exception of very long wavelengths in the RED spectrum)

Now... that said... we need to test and test again to make sure of it....LOL

If there is one thing I have learned... you never know until you test.

Regards,
Lightman

[critic2]

lightman.....your awesome. Great work!

[David_AVD]

They aren't flickering or PWM'ed at 56 kHz ( the modulated carrier freq of the detector)

This may be true of phase angle controlled AC LEDs, but don't be so sure with DC PWM. Some of those chip do PWM in the tens of KHz range.

I'm not saying it will be an issue, just don't discount it either.

[lightman]

I have been working on increasing the power on the Transmitter and am up tp 300 mW (was ~30 mW total). Range has increased, as expected.

I am driving 3 IR LED Emitters (the new TSAL7200) in series through a 2N2222A transistor driver with 2.2 ohm limiting resistor (5 volt Source). They are running at 180 mA each using carrier of 57 kHz... that equates to 100 mW/sr of radiant intensity each... or 300 mW/sr total. See attached schematic.

Attached photo shows the circuit as connected on protoboard. IR LED emitters are the clear ones. The Red LED is the visible transmission indicator. It twiddles on/off to signify when signal is being sent.

The transmitter is sitting on bench with the 3 IR LEDs aimed up at ceiling.... with a slight angle down (horizontal). Not the most efficient placement and aim.

Range has semi-proportionally increased. I am now out the door of my lab by bouncing off a 45 deg opened door... then down the hallway....for total of 33 ft (Updated for night time test indoors), all at 4800 baud. As you can see this is really inefficient and not direct Line of Sight, but we have been deluged with Thunderstorms and heavy Monsoons of late, so this inside test will have to do for now...LOL.

The real layout will be direct Line of Sight from transmitters to headset. I expect more range with that set up. I plan to use 3 transmitter stations positioned about 30 ft apart mounted on poles, in trees, or on the house. Each station will have a minimum of 3 IR LED emitters (300 mW). I can add more sets of 3 LEDs with another 2N2222 driver. Will play that by ear when I start installing outside.

Regards,
Lightman

[lightman]

Range Distance Testing:

Range/Distance Tests completed tonight with outstanding results! Max distance I could measure was 108 ft direct Line of Sight before I ran into obstacle (a house). Reception coverage area sprays 80% of the street in front of my house and well into neighbors yard. Here are the particulars....

Ran test in the evening with 3 IR LED emitter(TSAL7200) Transmitter sitting on my truck bed (~ 3 ft up) at edge of open garage. LEDs were arranged in basic linear (horizontal) pattern with a slight angular spread being driven at 180 mA each (300 mW/sr total). Headset Receiver used the TSOP34156 IR detector at 57 kHz carrier driving 2 GECE bulbs with 4 AAA batteries. IR Communication was at 4800 Baud. Vixen bus was running at 38.4 Baud.

Performed complete parametric mapping of the range (see scan attached). Coverage area exceeded my expectations and easily reaches across the street up to my neighbors garage and front door. Possibility for repeater aimed back the other way.

To read the map find the transmitter (TX) at bottom center of page (scan cut it off a little). It is transmitting up the page along radial lines, out from that point. Successful reception distance is marked with an "X". The dashed curved line indicates the edge of successful detection of signal. Anything inside the dashed lines is pick up by the headset. The bold black lines are the street, houses and driveways for reference.

Vixen drove the system using two sequences designed to stress the communication link at 50 msec rate. Moved headset out on radial line till the pattern started making errors... then I came back in... marked the spot and made distance measurement. Pattern of coverage appears to curve smoothly as expected in large oval shape.

I also moved the headset around to determine sensitivity to pitch and yaw. I was able to move the headset in pitch up 45 deg from horizontal and down by 45 deg before losing signal. Also got 45 deg of left and right azimuth before losing signal. This implies I will need two transmitters located at sides of yard aimed at 45 deg into the street. This will enhance pick up as kids move the headsets to and fro while dancing.

This is a line of sight communication system. If the transmitter gets blocked (lets say by my daughter leaving with a friend) the link shuts down. The instant they moved aside and the transmitter's line of sight was unobstucted again, the headset re-locked up and continued on with the sequence. When it loses lock the GECEs just hold their last command.... no weird flickering (This is a good thing).

I have included a couple of photos of the set up. SX based transmitter is sitting on top of beige box. Note the angle and spread of the IR LEDs... shooting over the SX uproc. The tailgate of my truck is about 3 ft or so off the ground with clear line of sight to street. You can also see Vixen running on netbook with Headset parked next to it.

This is terrific news!... With just a few IR LED emitters, a large swath of coverage is obtained... much larger than I expected. With a couple of these transmitter stations positioned at the sides of my yard, up high and aimed in and down a bit, they should easily cover the street, my neighbors yard, my other neighbor's yard.... etc. ....LOL.

Regards,
Lightman